INDEX

Explanations

instances of proper nouns referencing individuals or organizations

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 unden

-0.89

anwhile

-0.89

wcs

-0.84

ussion

-0.83

 livest

-0.82

 destro

-0.78

obyl

-0.77

 coincide

-0.77

 tradem

-0.75

quartered

-0.75

POSITIVE LOGITS

 Robot

0.89

Ack

0.88

 Olympia

0.86

 Obama

0.86

 Hyde

0.85

 Mund

0.85

 Spock

0.85

 Bezos

0.84

 Claus

0.84

 Rogers

0.84

Activations Density 0.037%

instances of proper nouns referencing individuals or organizations

No Comments

No Known Activations